Researcher — paper.dinq.me

2026 顶会论文

作者位次 #9/12NeurIPS 2026Poster

LAION-BVD: A 10-Million-Hour Open Video Dataset for Multimodal Pre-training

Andreas Hochlehnert · Marianna Nezhurina · Mehdi Cherti · Andrej Radonjic · Thaddäus Wiedemer · Christoph Schuhmann · Romain Beaumont · Wieland Brendel · Bernhard Schölkopf · A. Sophia Koepke · Jenia Jitsev · Matthias Bethge

多模态数据与预训练OpenReview

作者位次 #6/7NeurIPS 2026Poster· Deep Learning / Generative Models and Autoencoders

DiffRatio: Training One-Step Diffusion Models Without Teacher Supervision

Wenlin Chen · Mingtian Zhang · Jiajun He · Zijing Ou · José Miguel Hernández-Lobato · Bernhard Schölkopf · David Barber

多模态OpenReview

作者位次 #5/6NeurIPS 2026Poster

AM-Bench: A Unified Taxonomy and Evaluation Suite for Agentic Misalignment

Eric Zhang · Terry J. C. Zhang · Chijioke Ugwuanyi · Jerick Shi · Bernhard Schölkopf · Zhijing Jin

智能体与工具对齐与偏好评测与基准OpenReview

作者位次 #5/5NeurIPS 2026Poster

Intrinsically Interpretable Attention via Sparse Post-Training

Florent Draye · Anson Lei · Hsiao-Ru Pan · Ingmar Posner · Bernhard Schölkopf

可解释性强化学习与后训练OpenReview

作者位次 #4/5NeurIPS 2026Poster

Drifting Fields are not Conservative

Leonard T. Franz · Sebastian Hoffmann · Tim Weiland · Bernhard Schölkopf · Georg Martius

OpenReview

作者位次 #5/7NeurIPS 2026Poster

GT-HarmBench: Benchmarking AI Safety Risks Through the Lens of Game Theory

Pepijn Cobben · Xuanqiang A Huang · Thao Pham · Isabel Dahlgren · Bernhard Schölkopf · Terry J. C. Zhang · Zhijing Jin

安全评测与基准理论与学习动力学OpenReview

作者位次 #3/3NeurIPS 2026Poster

Perception for Action in Latent World Models

Petr Ivashkov · Randall Balestriero · Bernhard Schölkopf

OpenReview

作者位次 #7/10NeurIPS 2026Poster

Deriving Hyperparameter Scaling Laws via Modern Optimization Theory

Egor Shulgin · Jörg Franke · Dimitri von Rütte · Tianyue Zhang · Niccolò Ajroldi · Korbinian Pöppel · Bernhard Schölkopf · Aaron Klein · Peter Richtarik · Antonio Orvieto

数据与预训练理论与学习动力学OpenReview

作者位次 #10/11NeurIPS 2026Poster· Social Aspects / Fairness, Equity, Justice and Safety

SuperSycophantic: Stress-Testing Frontier LLMs from Single- to Multi-Turn Sycophancy

Terry J. C. Zhang · Oscar S Yasunaga · Wenyuan Jiang · Jessica Bo · Florent Draye · Aydin Javadov · Arth Singh · Lucille Muir · VICTORIA OLDEMBURGO DE MELLO · Bernhard Schölkopf · Zhijing Jin

社会与人本对齐与偏好安全OpenReview

作者位次 #2/3NeurIPS 2026Poster

PENEX: AdaBoost-Inspired Neural Network Regularization

Klaus-Rudolf Kladny · Bernhard Schölkopf · Michael Muehlebach

OpenReview

作者位次 #10/10NeurIPS 2026Poster· Reinforcement Learning / Everything Else

Learning POMDP World Models from Observations with Language-Model Priors

Valentin Six · Frederik Panse · Mathis Fajeau · Lancelot Da Costa · Mridul Sharma · Alfonso Amayuelas · Tim Xiao · David Hyland · Philipp Hennig · Bernhard Schölkopf

强化学习与后训练OpenReview

作者位次 #3/4NeurIPS 2026Poster

PVFormer: Proper Velocity Transformer for Stable and Scalable Hyperbolic Representation Learning

Xianglong Shi · Nicu Sebe · Bernhard Schölkopf · Ziheng Chen

可解释性OpenReview

作者位次 #3/3NeurIPS 2026Poster· General Machine Learning / Causality

Causality can systematically address the monsters under the bench(marks)

Felix Leeb · Zhijing Jin · Bernhard Schölkopf

OpenReview

作者位次 #3/4NeurIPS 2026Poster

Umbilic Multinomial Logistic Regression

Zihan Su · Nicu Sebe · Bernhard Schölkopf · Ziheng Chen

OpenReview

作者位次 #5/9NeurIPS 2026Poster· General Machine Learning / Everything Else

The Alien Space of Science: Sampling Coherent but Cognitively Unavailable Research Directions

Alejandro H. Artiles · Martin Weiss · Levin Brinkmann · Iyad Rahwan · Bernhard Schölkopf · Chris Pal · Hugo Larochelle · Anirudh Goyal · Nasim Rahaman

检索与知识OpenReview

Bernhard Schölkopf

获取完整研究档案

联系方式

2026 顶会论文

LAION-BVD: A 10-Million-Hour Open Video Dataset for Multimodal Pre-training

DiffRatio: Training One-Step Diffusion Models Without Teacher Supervision

AM-Bench: A Unified Taxonomy and Evaluation Suite for Agentic Misalignment

Intrinsically Interpretable Attention via Sparse Post-Training

Drifting Fields are not Conservative

GT-HarmBench: Benchmarking AI Safety Risks Through the Lens of Game Theory

Perception for Action in Latent World Models

Deriving Hyperparameter Scaling Laws via Modern Optimization Theory

SuperSycophantic: Stress-Testing Frontier LLMs from Single- to Multi-Turn Sycophancy

PENEX: AdaBoost-Inspired Neural Network Regularization

Learning POMDP World Models from Observations with Language-Model Priors

PVFormer: Proper Velocity Transformer for Stable and Scalable Hyperbolic Representation Learning

Causality can systematically address the monsters under the bench(marks)

Umbilic Multinomial Logistic Regression

The Alien Space of Science: Sampling Coherent but Cognitively Unavailable Research Directions

常见合作者